A Method for Hyperspectral Images Classification using Spectral Correlation

被引：0

作者：

Jing, Luo ^{[1
,2
]}

Fei, Hu ^{[1
,2
]}

Li Yun-lei ^{[1
,2
]}

Yue, Liu ^{[1
,2
]}

机构：

[1] Key Lab Adv Elect Engn & Energy Technol, Tianjin 300387, Peoples R China

[2] Tianjin Polytech Univ, Coll Elect Engn & Automat, Tianjin 300387, Peoples R China

来源：

2018 37TH CHINESE CONTROL CONFERENCE (CCC) | 2018年

基金：

中国国家自然科学基金;

关键词：

hyperspectral images classification; band correlation coefficient selection (BCCS); spectral correlation; Spectral Correlation(SC); ALGORITHM;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

To reduce the effects of information redundancy caused by noise and band correlation on the hyperspectral image analysis, a method for hyperspectral images classification using Spectral Correlation(HCSC) is proposed. Firstly, the hyperspectral images are preprocessed. Secondly, band selection algorithm based on correlation coefficient(BCCS) is proposed to reduce the information redundancy caused by adjacent bands. Thirdly, The spatial pixel purity index(SPPI) is used to extract the pure pixel and end-member extraction is accomplished. The experiments have been done on AVIRIS database. The experimental results have shown the proposed algorithm(HCSC) can improve the efficiency in the pure pixel extraction and classification, and the classification accuracy is 86.4%.

引用

页码：9067 / 9072

页数：6